ABSTRACT

advertisement
EXTENDED ABSTRACT -- OC/EC Workshop, Durango, CO, March 3-5, 2003
How Does The Thermal-Optical Transmission Method Behave Optically?
A Focus On The Apparent Specific Absorption Cross Section
J.M. Conny, Surface and Microanalysis Science Division, National Institute of Standards and
Technology, 100 Bureau Drive Stop 8371, Gaithersburg, MD 20899 (joseph.conny@nist.gov)
Thermal-optical analysis (TOA) for measuring elemental carbon (EC) in particulate matter (PM)
relies on changes in the optical behavior of PM carbon during thermal desorption to indicate
when organic carbon (OC) mass analytically ‘separates’ from EC mass. An emphasis on TOA’s
optical behavior during thermal desorption presents it as a method for measuring black carbon
(BC), and thus allows for a definitional link between EC and BC. With this link in mind, the
major advantage of TOA for measuring EC over purely optical methods is that no prior
knowledge of PM optical absorptivity is required.
In order to view thermal-optical transmission (TOT) as an accurate method for measuring BC
(hence, EC), there are three assumptions of the constancy of absorptivity during analysis:
1. Absorptivity of carbonaceous PM remains constant prior to OC pyrolysis, i.e., no substantive
change to the particle refractive index occurs that would affect absorptivity.
2. Pyrolyzed OC absorbs at the analytical wavelength (hence, chars), and absorptivity remains
constant after its formation within a high-temperature step in the non-oxidizing (He) phase.
3. Pyrolyzed OC has the same absorptivity as BC that is native to sample.
These assumptions were tested by determining the variation in TOT absorptivity during both the
He phase and the subsequent oxidizing (O2-He) phase for three types of samples: NIST SRM
1649a (urban dust) fine fraction, fine indoor air particles, and smoldering forest-fire emissions.
Absorptivity was calculated as the apparent specific absorption cross section following the
approach of Petzold et al. [1]. Analogous to absorbance at the analytical wavelength (670 nm),
the temperature-dependent attenuation was calculated as
 I(T) 

ATN(T)  100 ln 
I
0


(1)
I(T) is the temperature-dependent laser response during the course of analysis, and I0 is the laser
response following the desorption of carbon, i.e., at the end of the TOT run, and corrected to the
initial operating temperature (50 ºC).
ATN(T) can be described as the sum of two components: attenuation due to absorbance and
attenuation due to scattering. ATN(T) can also be described as the sum of the temperaturedependent attenuation due to the carbonaceous material (ATNC(T)) and the temperaturedependent attenuation due to non-carbonaceous material (ATNnon-C(T)):
ATN (T )  ATN C (T )  ATN nonC (T )
(2)
It is important to note that I0 (eq. 1) accounts for the laser response from non-carbonacous
material at 50 ºC. Therefore, neither ATN in eq. 1 nor ATNnon-C in eq. 2 includes the attenuation
due to the non-carbonacous component at 50 ºC.
Scattering is likely the dominant contributor to ATNnon-C at elevated temperatures. Absorbance
and scattering both contribute to ATNC. However, in this case absorbance is likely the major
contributor. Therefore, ATNC(T) is equal to the product of the specific carbon mass loading (SC,
g cm-2) and the apparent specific absorption cross section (σC, m2 g-1):
ATNC (T )   C (T ) SC
(3)
The figure below shows ATN(T) and σC for the lab air sample using the following desorption
protocol: He step 1, 250 ºC for 120 s; He step 2, 450 ºC for 168 s; He step 3, 650 ºC for 222 s;
He step 4, 850 ºC for 270 s; O2-He step 1, 550 ºC for 60 s, O2-He step 2, 700 ºC for 60 s; O2-He
step 3, 850 ºC for 60 s; O2-He step 4, 900 ºC for 120 s.
For lab air particles (as well as the other two samples) the apparent specific absorption cross
section is obviously not constant throughout TOT. From the figure, assumptions 2 and 3 in
particular, are violated. For example, σC does not remain constant after the formation of char
during step 4 in He (assumption 2). In addition, σC values at the end of the He phase and at the
split point are not equivalent (assumption 3). However, if step 4 in He is shortened from 270 s to
150 s, the split point shifts toward the right to where σC is close to its value at the end of the He
phase, thus satisfying assumption 3. These results suggest that analysis of TOT absorptivity in
this manner along with response surface methods [2] may lead to sample-specific TOT protocols
that are optimized for accuracy in measuring EC (as BC).
[1] Petzold, A., C. Kopp, and R. Niessner, Atmos. Envir., 31, 661-672 (1997).
[2] Conny, J.M., D.B. Klinedinst, S.A. Wight, J.L. Paulsen, Aerosol. Sci. Technol. (submitted).
Download